by methods 1 and 3 of the molecular modeling core. All projects have specific aims that can be aided by methods 2 and 3 of the core. QSAR and chemoinformatic studies of the screened compounds, can provide information on molecular features such as lipophilicity, orientation of hydrogen bonding groups, and molecular volumes. Such studies will allow the elucidation of common features for active and inactive compounds, the prediction of biological activities including properties such as membrane accumulation. There will be 3 types of computational chemistry in the molecular modeling effort: Modeling Methods (1) Comparative (Homology) Models of rat a1p2Y2 pentameric GABA-A receptor. PHS 398/2590 (Rev.09/04, Reissued 4/2006) Page o PT 1 Continuation Format Page Principal Investigator/Program Director (Last, First, Middle): Steinbach, Joseph Henry (2) QSAR studies of previously utilized compounds (3) Docking studies of compounds to receptor models These molecular modeling studies will be carried out in the laboratory of Dr. David Reichert in our Department of Radiology. The Reichert lab's research is primarily focused on the computer aided molecular design of in vivo imaging agents. As part of this work they have developed expertise in the modeling of steroids and steroid analogs for imaging estrogen receptor expression. In order to validate methodologies that they believed could be successfully applied to imaging agents, they first developed computer models capable of predicting the binding affinities of known ligands for both known isoforms of the estrogen receptor (a and P). This work has been published as two papers in the Journal of Computer-Aided Molecular Design, and the Journal of Molecular Graphics and Modelling [43,44]. Comparative modeling - The goals proposed for this core require structural models of the GABA-A receptor. Although the three dimensional structure of many enzymes and receptors are still unknown, the number of experimentally determined structures is rapidly growing. In 1995, the October release of the Protein Data Bank (PDB) [7] had 3,821 structures [34]. As of September 2006, this number had grown to 38,620. Unfortunately, the GABA-A receptor is not one of the known structures. Two related structures are known and have been used by several groups to build comparative models of the GABA-A receptor, these are the acetylcholine-binding protein (AChBP) from Lymnaea stagnalis (PDB ID 119B)and the nicotinic acetylcholine receptor (nAChR) from Torpedo marmorata (PDB ID 2BG9). Recent examples of the development of comparative models of the GABA-A receptor are from Trudell and Bertaccini [40], Ernst et at [15], and Campagna-Slater and Weaver [9]. Of particular utility is the work of Ernst et a\, who developed a model of the rat a<\$2\2 pentameric receptor [15]. The development of comparative models for the GABA-A receptor is a non-trivial task, primarily due to the low sequence identity between members of the "cys-loop" family [14]. In general, with a sequence identity >60% pairwise sequence alignment is quite accurate, this falls off quickly with a sequence identity < 40%. The sequence identity between AChBP and "cys-loop" extracellular domains ranges from 15-30%. The situation for the membrane spanning regions is even worse, the sequence identity for the rat cti subunit to the nAChr subunits ranges from 19- 21%. Despite this fact, reasonable alignments can be produced using conserved positions; QSAR and CoMFA - Quantitative structure-activity relationships (QSAR) have become a common tool in the field of molecular modeling since their introduction [20]. Indeed they have found application in both the prediction of biological activity and more recently in the prediction of the Absorption, Distribution, Metabolism, Excretion and Toxicological (ADME/tox) properties of organic drug-like compounds [4,17-19,25]. A related technique CoMFA (Comparative Molecular Field Analysis) has been utilized extensively to study the relationship between three-dimensional molecular information such as steric and electrostatic fields and biological activity [12,21,42]. CoMFA is based on the premise that the pharmacophoric elements which are responsible for the biological activity of a compound will be represented in the calculated steric and electrostatic fields of the compound. By studying a series of compounds, called the training set, consisting of compounds with good, medium and poor bioactivity for a specific protein target it is possible to extrapolate a three-dimensional pharmacophoric model that explains the observed bioactivity. Indeed this model suggests how the steric and electrostatic fields might be manipulated to produce a novel compound with enhanced bioactivity. One requirement of CoMFA is that the compounds in the training set be aligned against each other so that the overlap of the pharmacophoric elements responsible for producing a biological response is maximized. In cases where the ligands are very diverse in structure or have several possible modes of binding, developing the alignment can be problematic. In cases where the crystal structure of the target protein complexed to a ligand has been resolved, the structure of the docked ligand can be used as a template. However, even in this advantageous case it is difficult to deal with compounds in the training set which might have multiple protein binding conformations while maintaining a high pharmacophoric overlap with the template compound. A new approach to this problem is to use a docking program capable of predicting the most favorable conformation of the bound ligand without introducing any human bias. Molecular Docking - The objective of molecular docking is to obtain the lowest free energy structure forthe ligand - receptor complex. As stated by Kuntz in 1994, the docking problem can be divided into three components [26]. The first is the representation of the binding